Abstract
Cytoplasmic bacteria of the genus Wolbachia are best known as the cause of cytoplasmic incompatibility (CI): many uninfected eggs fertilized by Wolbachia-modified sperm from infected males die as embryos. In contrast, eggs of infected females rescue modified sperm and develop normally. Although Wolbachia cause CI in at least five insect orders, the mechanism of CI remains poorly understood. Here I test whether the target of Wolbachia-induced sperm modification is the male pronucleus (e.g., DNA or pronuclear proteins) or some extranuclear factor from the sperm required for embryonic development (e.g., the paternal centrosome). I distinguish between these hypotheses by crossing gynogenetic Drosophila melanogaster females to infected males. Gynogenetic females produce diploid eggs whose normal development requires no male pronucleus but still depends on extranuclear paternal factors. I show that when gynogenetic females are crossed to infected males, uniparental progeny with maternally derived chromosomes result. This finding shows that Wolbachia impair the male pronucleus but no extranuclear component of the sperm.
WOLBACHIA comprises a group of maternally transmitted cytoplasmic bacteria that have a variety of reproductive effects in arthropods (O'Neillet al. 1997; Werren 1997). These bacteria have been implicated as the cause of induced parthenogenesis in haplodiploids (Stouthameret al. 1990), feminization of genetic males in isopods (Roussetet al. 1992), and, most notably, cytoplasmic incompatibility (CI; Yen and Barr 1971; Hoffmann and Turelli 1997). CI results when the sperm of Wolbachia-infected males fertilize eggs from uninfected females, causing embryonic death. Crosses between infected males and infected females produce no developmental anomalies. Antibiotic treatment of infected individuals simultaneously cures the Wolbachia infection and the associated reproductive effects (Wright and Barr 1981). The ability of Wolbachia strains to induce CI after experimental transfer between phylogenetically distant hosts (i.e., between the mosquito Aedes albopictus and the fruit fly Drosophila simulans) suggests that the microbes disrupt an evolutionarily conserved target (Braiget al. 1994). But the mechanism by which Wolbachia cause CI remains unknown.
Although Wolbachia are abundant in the testes of infected males, they are not physically associated with mature sperm (Binnington and Hoffmann 1989; Bressac and Rousset 1993). Instead the bacteria are shed with the cytoplasm during individualization in spermatogenesis. Wolbachia do not therefore cause CI directly, but modify developing sperm, which then transmit the CI-inducing effects to eggs. Cytological studies of eggs from incompatible crosses reveal early mitotic defects and paternal chromosome loss following fertilization in the mosquitoes Culex pipiens (Jost 1970) and Aedes polynesiensis (Wright and Barr 1981), the wasp Nasonia vitripennis (Ryan and Saul 1968; Reed and Werren 1995), and the fruit fly D. simulans (Callaini et al. 1996, 1997; Lassy and Karr 1996). Haploid development results in production of male progeny in the haplodiploid Nasonia (Breeuwer and Werren 1990) but is lethal in mosquitoes and Drosophila. Only the presence of Wolbachia (or Wolbachia-derived products) in the egg cytoplasm can rescue such modified sperm.
To cause CI, Wolbachia must modify nuclear and/or extranuclear components of the sperm. For example, modification of paternal chromosomes—either DNA or paternal DNA-packaging proteins—during spermatogenesis might later disrupt the condensation cycle of the male pronucleus and/or karyogamy after fertilization. Alternatively, Wolbachia might modify extranuclear factors of sperm that are essential for embryonic development but unrelated to processing of the male pronucleus. There are several examples of such paternal factors. In most animals, for instance, the paternal centrosome is essential: centrosome elements of the sperm basal body must combine with those of the maternal centrosome to form the zygotic microtubule organizing center (MTOC). The MTOCs replicate and orchestrate assembly of the spindles needed for pronuclear apposition and segregation of chromosomes during mitosis (Schatten 1994). In addition to the centrosome, several essential paternal proteins have been identified in Drosophila (reviewed in Karr 1996; Fitchet al. 1998) and Caenorhabditis elegans (Browning and Strome 1996). Moreover, because the entire sperm enters the egg in many Drosophila species, it has been suggested that the sperm tail itself plays a critical role in the early embryo (Karr 1996). All such extranuclear paternal contributions represent potential targets for Wolbachia.
In fact, cytological work in D. simulans has uncovered both abnormal paternal chromosome behavior and irregular centrosome-mediated microtubule processes in embryos from incompatible crosses (Callainiet al. 1996; Lassy and Karr 1996). Two unpublished studies also found differences in sperm proteins between infected and uninfected Drosophila males (cited in Karr 1996; T. Sasaki cited in Wilkinson 1998). While suggestive, these findings do not conclusively identify the site of Wolbachia's action. Distinguishing between the nuclear vs. extranuclear possibilities thus represents an important step toward understanding the mechanism of CI.
Here I present genetic results that distinguish these possibilities, using a gynogenetic stock of D. melanogaster (Fuyama 1984, 1986a,b). Gynogenesis is like parthenogenesis in that diploid zygotes inheriting all chromosomes from their mother can develop without a genetic contribution from males. However, as shown in crosses below, even though gynogenetic diploid eggs do not require paternal chromosomes, they do require extranuclear factors from sperm—physical penetration of the egg alone is insufficient to initiate development. We can therefore ask whether diploid gynogenetic eggs can develop using the sperm of Wolbachia-infected males: If Wolbachia disrupt paternal chromosomes only, diploid gynogenetic eggs should develop; if, however, Wolbachia disrupt any extranuclear paternal factors required for development, diploid gynogenetic eggs should not develop (see Figure 1). Thus, by crossing uninfected gynogenetic females to infected males, I can determine whether Wolbachia disrupt the male pronucleus or essential extranuclear paternal factors.
MATERIALS AND METHODS
Stocks: Fly stocks were kindly provided by Drs. Y. Fuyama [w; gyn-2; gyn-3 and ms(3)K81], K. Fitch [ms(3)snky], S. O'Neill [Wolbachia-infected D. melanogaster Canton-S], A. Hoffmann [D. simulans Riverside (DSR)], and M. Turelli [D. simulans Watsonville (DSW)].
Infection status and CI: Tetracycline curing of Wolbachia infections was carried out as described by Hoffmann et al. (1986). Flies were bred on standard medium with 0.3% tetracycline concentration for at least three generations. “DSRT” and “Canton-ST” refer to tetracycline-cured DSR and Canton-S flies.
The infection status of all stocks was confirmed by PCR using primers specific for Wolbachia pipientis and the reaction conditions described in O'Neill et al. (1992). An ~900-bp fragment from the bacterial 16S rRNA gene was amplified from Wolbachia-infected strains, while no product was obtained from uninfected strains.
The cross between gynogenetic w; gyn-2; gyn-3 females and Wolbachia-infected males. w; gyn-2; gyn-3 females produce both diploid eggs and haploid eggs; infected males produce both modified sperm and unmodified sperm. Cells indicate the possible genotypes of daughters. If CI is caused by modification of the male pronucleus only, white-eyed uniparental daughters will be seen (top left). If, on the other hand, CI is caused by modification of extranuclear paternal factors, these white-eyed uniparental daughters will not be seen—they instead will die as embryos.
Levels of CI were measured for both intraspecific and interspecific crosses at 25°. Single pairs of flies were set up in vials with standard medium for 24 hr. Females were then transferred to vials containing small spoons with grape juice-colored medium coated with a live yeast suspension (Hoffmannet al. 1986). Females were transferred every 24 hr for several days. The percentage egg hatch was scored 28 hr after females were removed from a vial. To control for background egg mortality (i.e., mortality independent of Wolbachia), I calculated corrected CI values (CIcorr; Poinsotet al. 1998) as the rate of egg mortality from incompatible crosses (CIobs) minus the rate of egg mortality observed in compatible crosses (CCM) between two uninfected individuals: CIcorr = [(CIobs − CCM)/(100 − CCM)].
Crosses: The gynogenetic D. melanogaster stock, w; gyn-2; gyn-3, is described in detail by Fuyama (1986b). w; gyn-2; gyn-3 females can produce both biparental and uniparental progeny. Expression of a recessive visible mutation (in this case white) carried by the mother allows a simple check that the chromosomes of gynogenetically produced uniparental progeny are maternally derived, as all males used in this study had wild-type eye color, w+.
I performed two versions of the experimental test, one using crosses between species and one using crosses within species. Because infected strains of D. melanogaster show only low levels of CI (e.g., Hoffmannet al. 1994; A. Hoffmann, T. Karr, M. Turelli, F. Rousset, M. Solignac, personal communications), I initially crossed w; gyn-2; gyn-3 melanogaster females to males of another species, the infected Riverside strain of D. simulans. DSR shows high levels of CI within D. simulans (Hoffmannet al. 1986). Using species hybrids does not affect the level of CI: uninfected D. melanogaster females crossed to infected D. simulans males show levels of CI comparable to those seen within D. simulans (M. Green, personal communication; and below). Ten w; gyn-2; gyn-3 virgin females were mass mated to 20 3- to 5-day-old virgin D. simulans males. Males were removed from vials when there was evidence of fertilization (i.e., dead eggs or larvae), usually within 2–3 days. D. melanogaster females crossed to D. simulans males produce hybrid daughters only; male hybrids die at the larval-pupal transition (Sturtevant 1920). Lethality of hybrid males in no way affects the present results, as tests of gynogenesis among w; gyn-2; gyn-3 females involve scoring uniparental w/w daughters.
I repeated the experiment within species. w; gyn-2; gyn-3 females were crossed to an infected strain of D. melanogaster created by introgressing Canton-S chromosomes into the infected cytoplasm of a y w stock (S. O'Neill, personal communication). Because CI is weak in D. melanogaster, I tried to increase its expressivity by using a slightly different crossing design from that used above. I set up individual w; gyn-2; gyn-3 virgin females in vials with 3–5 1-day-old Canton-S virgin males and observed all crosses until copulation occurred, as CI levels decrease with male age (Hoffmannet al. 1986) and repeated copulation (Karret al. 1998). Each female was then aspirated and placed in a fresh vial for oviposition. Females were transferred to new vials every 3–4 days. The eye color and sex of all progeny were scored.
When haploid w; gyn-2; gyn-3 eggs are fertilized by unmodified w+ sperm, red-eyed diploid daughters result (w+/w; Figure 1). When these eggs are fertilized by Wolbachia-modified sperm, CI results (Figure 1). However, ~24% of w; gyn-2; gyn-3 eggs are diploid (see Table 1 in Fuyama 1986b). When diploid w; gyn-2; gyn-3 eggs are fertilized by umodified w+ sperm, triploid red-eyed daughters result (w+/w/w). When these eggs are fertilized by Wolbachia-modified sperm, and if only the paternal chromosomes are disrupted by Wolbachia, production of uniparental white-eyed daughters should result (Figure 1). We can roughly approximate the expected percentage of uniparental white-eyed daughters as
RESULTS AND DISCUSSION
Properties of the gyn-2; gyn-3 stock: Before performing the key experiment, it is important to characterize the properties of gyn-2; gyn-3 reproduction and to characterize the infection and CI status of the stocks employed.
First, both haploid and diploid eggs of uninfected w; gyn-2; gyn-3 females use the nuclear and extranuclear contributions of unmodified w+ sperm: w; gyn-2; gyn-3 females crossed to uninfected Watsonville (DSW) males produce only red-eyed daughters (Table 1, line 1; see Table 2 for infection status). Because approximately one-third of w; gyn-2; gyn-3 eggs are diploid (Fuyama 1986b), some red-eyed daughters are w+/w/w triploids.
Next, two facts must be established: (1) diploid w; gyn-2; gyn-3 eggs do not require a paternal nuclear contribution; and (2) diploid w; gyn-2; gyn-3 eggs do require an extranuclear paternal contribution. I confirmed that w; gyn-2; gyn-3 females can reproduce gynogenetically by crossing them to homozygous ms(3)K81 males (Fuyama 1984, 1986b). While functional in all other respects, ms(3)K81 sperm cannot deliver the paternal pronucleus. Thus, if w; gyn-2; gyn-3 females reproduce gynogenetically, the cross of w; gyn-2; gyn-3 females to ms(3)K81 males should produce abundant uniparental white-eyed daughters. This is precisely what occurred: virtually all progeny (99.8%) were white-eyed uniparental daughters (Table 1, line 2). In contrast, nongynogenetic Oregon-R females produced only dead eggs when crossed to ms(3)K81 males (Table 1, line 3). Diploid w; gyn-2; gyn-3 eggs do not therefore require a paternal nuclear contribution. The few sons produced (6.7%) in the first cross were sterile XO males (confirmed by testes dissections and failure to produce progeny) resulting from nondisjuction in one of the two egg pronuclei that fuse to restore diploidy in uniparental progeny [in a similar cross Fuyama (1986b) found ~2% XO males].
The ability of diploid w; gyn-2; gyn-3 eggs to use ms(3)K81 sperm shows that they do not need a nuclear contribution from males. But the fact that these eggs never develop without fertilization shows that they require something from the sperm. To test whether sperm penetration alone is sufficient to stimulate diploid w; gyn-2; gyn-3 egg development, I crossed w; gyn-2; gyn-3 females to D. melanogaster males homozygous for the paternal effect lethal mutation ms(3)snky. The plasma membrane of snky sperm fails to break down after penetration, trapping the nuclear and extranuclear paternal contributions within the membrane (Fitch and Wakimoto 1998; Fitchet al. 1998). If diploid w; gyn-2; gyn-3 eggs only require sperm penetration, uniparental white-eyed daughters should appear. Instead, the cross of w; gyn-2; gyn-3 females to ms(3)snky males produced only a single biparental daughter from thousands of eggs (Table 1, line 4). Thus, sperm penetration alone is not sufficient to initiate development—diploid w; gyn-2; gyn-3 eggs require essential extranuclear factors from sperm. The single escaper reflects ms(3)snky's known slight leakiness (Fitch and Wakimoto 1998; see also the single male produced in the control Oregon-R cross, Table 1, line 5).
Gynogenetic reproduction
Infection status as determined by PCR assay
Infection status and CI: Table 2 gives the infection status of all stocks. Table 2 also shows that Wolbachia infections were successfully cured by the tetracycline treatment.
Within-species levels of CI (Table 3) were similar to those in previous reports (e.g., Hoffmannet al. 1986; Bourtziset al. 1998). In D. simulans, DSR males cause strong CI when crossed to DSRT females (CIcorr = 79.9%). As expected for D. melanogaster, Canton-S males cause weak CI when crossed to cured Canton-ST females (CIcorr = 25.3%).
The level of CI induced by DSR males between species (Table 3, line 10) was similar to that seen within species: uninfected D. melanogaster Oregon-R females mated to DSR males showed CIcorr = 87.0%. Uninfected D. melanogaster Oregon-R females mated to cured DSRT males showed egg hatch rates similar to compatible crosses within species (Table 3, line 11). Species crosses between D. melanogaster w; gyn-2; gyn-3 females and DSR males should therefore reflect normal DSR levels of CI.
Test of Wolbachia target: We now turn to the critical experiment. The crosses above show that w; gyn-2; gyn-3 eggs use the paternal chromosomes and extranuclear factors of unmodified wild-type sperm (Table 1, line 1). They also show that diploid w; gyn-2; gyn-3 eggs do not require paternal chromosomes (Table 1, line 2) but do require extranuclear factors from the sperm (Table 1, line 4). Given this, we can make two predictions: If Wolbachia impair the paternal chromosomes only, uniparental diploid white-eyed daughters should appear when w; gyn-2; gyn-3 females are crossed to infected males. If, on the other hand, Wolbachia impair an essential extranuclear paternal factor, uniparental white-eyed daughters should not appear.
Cytoplasmic incompatibility relationships among strains
Crosses between gynogenetic females and infected males
When w; gyn-2; gyn-3 females are crossed to infected DSR males, many uniparental daughters are produced: 52% of daughters are white-eyed (Table 4, line 1). This result is not an artifact of the species cross, as similar results are obtained in the within-species test: when w; gyn-2; gyn-3 females are crossed to infected Canton-S males, 16.1% white-eyed daughters appear (Table 4, line 2).
As expected, control crosses of w; gyn-2; gyn-3 females to tetracycline-cured DSRT males failed to produce any white-eyed daughters (Table 4, line 3). Similarly, w; gyn-2; gyn-3 females crossed to cured D. melanogaster Canton-ST males produce only rare escaper (<1%) white-eyed daughters (Table 4, line 4). Tetracycline treatment of infected males thus cures Wolbachia infection and simultaneously eliminates production of uniparental progeny.
The production of uniparental daughters in crosses to infected males definitively shows that w; gyn-2; gyn-3 eggs successfully use the extranuclear components of Wolbachia-modified sperm but not the male pronucleus. Wolbachia do not, therefore, impair essential extranuclear components of the sperm.
It is worth noting that the levels of gynogenetic reproduction induced by DSR and Canton-S males are proportional to the levels of CI induced by males of these strains. This reflects the fact that both phenotypes—percentage uniparental progeny and percentage of unhatched eggs (CI)—are largely determined by the percentage of sperm modified. For example, given the 25.3% sperm modification rate of the Canton-S Wolbachia strain and the 24% diploid egg production of w; gyn-2; gyn-3 females, we expect ~7.5% white-eyed daughters (= (0.25 × 0.24)/[(0.25 × 0.24) + 0.75(0.24+ 0.76)]). We observe 16% white-eyed daughters (Table 4, line 2). Similarly, given the 79.9–87.0% sperm modification rate of the DSR strain, we expect 48.8–61.6% white-eyed daughters. We observe 52% (Table 4, line 1). Finally, for crosses involving males from uninfected strains, no sperm modification occurs and virtually no uniparental daughters are produced.
Concluding remarks: The genetic tests performed here support the results of cytological studies in Drosophila and Nasonia suggesting that CI is caused by disruption of paternal chromosome processing in uninfected eggs (O'Neill and Karr 1990; Reed and Werren 1995; Callaini et al. 1996, 1997;Lassy and Karr 1996). Similarly, the fact that CI in the haplodiploid wasp Nasonia results in haploid males shows that the paternal genome is affected. However, none of these earlier findings ruled out the possibility that Wolbachia also impair extranuclear sperm components. In fact, previous cytological work showed that—in addition to chromosomal mitotic defects—abnormal spindle structures and supernumerary centrosomes appear in incompatible crosses (Callainiet al. 1996; Lassy and Karr 1996). The present results, however, rule out the possibility that Wolbachia critically impair the paternal centrosome, the sperm tail, or any other essential extranuclear paternal factors not related to paternal chromosome processing. Instead, Wolbachia critically impair the male pronucleus.
The present results do, however, suffer from at least one limitation: they tell us little about the cellular and molecular mechanisms of CI. The questions of how the male pronucleus is modified and how Wolbachia in the egg rescue this modification still remain. As noted by Karr and colleagues (O'Neill and Karr 1990; Karr 1996; Lassy and Karr 1996), the embryonic lethal phenotypes caused by ms(3)K81 sperm and Wolbachia-modified sperm are strikingly similar. The present work shows that both lesions can induce gynogenetic reproduction in gyn-2; gyn-3 females. Based on their genetic analysis of ms(3)K81, Yasuda et al. (1995) conclude that the K81 protein is likely required for one of three steps in postfertilization chromosome remodeling: decondensation of the sperm pronucleus, replication of the male pronucleus, or recondensation of the male pronucleus corresponding to its acquisition of maternally derived chromatin packaging proteins. Wolbachia might similarly interfere with any one of these critical steps.
Acknowledgments
I thank Andrea “Texas” Betancourt, Seth Bordenstein, Jerry Coyne, John Jaenike, Corbin Jones, Tim Karr, Michael Turelli, Jack Werren, and especially Allen Orr for helpful discussion and comments. This work was supported by grants from the National Institutes of Health (GM-51932) and the David and Lucile Packard Foundation to H. A. Orr and an Ernst Caspari fellowship to the author.
Footnotes
-
Communicating editor: A. G. Clark
- Received June 25, 1999.
- Accepted November 10, 1999.
- Copyright © 2000 by the Genetics Society of America
